As part of CSCE436 Human Computer Interaction, our team was tasked with solving an HCI problem. Throughout my time on the internet, I’ve gotten tired of frequently seeing aggressive terms of service agreements. While we can’t make TOS agreeable, we can make them easier to read.
Using both LLMs (ChatGPT API) and a heuristic rating system, our team created a Chrome extension that parses and rates TOS agreements. It will automatically detect if you are on a TOS page and, if so, will offer to analyze the page. Doing so will return a dashboard full of key metrics about the TOS.
This program was created to display information about TOS in descending levels of abstraction- it would give more in-depth summaries the more you look around the extension. This was to allow the user to grasp the general rating of the website or, if they were so inclined, to read more.
Our project saw a 5% increase in user accruacy when taking quizzes about read TOS with our extension, a 17% increase in inferencing, 29% decrease in time taken to parse TOS, and an 80% return rate (they would use this extension again).
Gallery
